Lipoprotein(a) and Cardiovascular Outcomes in Patients With Coronary Artery Disease and Prediabetes or Diabetes

Lipoprotein(a) [Lp(a)] is an LDL-like particle consisting of a moiety of apolipoprotein(a) bounding covalently to apolipoprotein B-100 (1,2). Plasma Lp(a) induces proatherogenic effects via LDL moiety and prothrombotic effects by the plasminogen-like apolipoprotein(a) (3,4). Strong evidence in epidemiological, genetic, and prospective cohort studies verified that circulating Lp(a) levels were associated with the presence of cardiovascular disease (CVD) (5–7). In the Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High Triglycerides: Impact on Global Health Outcomes (AIM-HIGH) study, Lp(a) was also associated with increased cardiovascular event (CVE) risk in patients with established CVD and remains predictive for CVE risk at LDL cholesterol (LDL-C) levels <1.8 mmol/L (8).

Type 2 diabetes mellitus (T2DM) is considered as a CVD risk equivalent in many guidelines (9). These guidelines also recommend lipid targets for T2DM patients as LDL-C <1.8 mmol/L and non-HDL cholesterol (non-HDL-C) <2.6 mmol/L, which is the same as for patients with established coronary artery disease (CAD) (10,11). Although patients with T2DM did not exhibit higher levels of Lp(a), results from the Biomarker for Cardiovascular Risk Assessment across Europe (BiomarCaRE) consortium indicated that elevated Lp(a) was robustly associated with an increased risk for first ever CVD in individuals with diabetes mellitus (DM) (12,13). Pre–diabetes mellitus (pre-DM) is an intermediate state between normal glucose regulation (NGR) and DM, with a 35.7% morbidity rate in China (14). Recently, Saeed et al. (12) reported that elevated Lp(a) levels in Caucasian individuals with DM or pre-DM were positively related to CVD risk in a primary prevention study. Our previous studies suggested that Lp(a) levels were strongly associated with the presence and severity of stable CAD in individuals with DM and patients with familial hypercholesterolemia (15,16). In the current study, we try to examine the relation of elevated Lp(a) levels to the risk of CVEs in stable CAD patients with different glucose metabolism on optimal medications.

Our study complied with the Declaration of Helsinki and was approved by the hospital’s ethics review board (Fu Wai Hospital, National Center for Cardiovascular Diseases). Informed written consents were obtained from all patients enrolled in this study.

As described in the flowchart (Supplementary Fig. 1), from March 2011 to February 2015, 6,811 patients scheduled for coronary angiography because of angina-like chest pain and/or positive treadmill exercise test or clinically suspected CAD were recruited from three medical centers. Among these patients, 569 were excluded because they did not have angiography-proven CAD (coronary stenosis ≥50% of at least one coronary artery). Other patients were excluded for the following reasons: acute coronary syndrome, previous myocardial infarction (MI), percutaneous coronary artery intervention or bypass grafting, heart failure, severe liver and/or renal insufficiency, thyroid dysfunction, systematic inflammatory disease, and malignant disease. Patients were followed up at 6-month intervals by means of direct interview or telephone. Trained nurses or physicians who were blinded to the clinical data completed the interview. The primary end points (CVEs) were cardiovascular mortality, nonfatal MI, and stroke. Nonfatal MI was diagnosed as positive cardiac troponins along with typical chest pain or typical electrocardiogram serial changes. Stroke was diagnosed by the presence of typical symptoms and imaging.

DM was diagnosed by fasting plasma glucose ≥7.0 mmol/L, the 2-h plasma glucose of the oral glucose tolerance test ≥11.1 mmol/L, or current use of hypoglycemic drugs or insulin. Pre-DM was diagnosed in participants who had no self-reported DM or hypoglycemic therapies but had a fasting plasma glucose ranging from 5.6 to 6.9 mmol/L, 2-h glucose ranging from 7.8 to 11.0 mmol/L, or hemoglobin A_1c (HbA_1c) level ranging from 5.7 to 6.4% (17). Patients who were without DM or pre-DM were defined as having normal glucose regulation (NGR). Hypertension was defined as self-reported hypertension, currently taking antihypertensive drugs, or recorded systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg three or more consecutive times. Information regarding other disease, family history, and prior therapy of every patient was collected from self-reported medical history.

Blood samples were obtained from each patient from the cubital vein after at least 12 h of fasting. Concentrations of total cholesterol (TC), triglyceride (TG), LDL-C, and HDL-C were measured using an automatic biochemistry analyzer (7150; Hitachi, Tokyo, Japan) in an enzymatic assay. Lp(a) was determined by immunoturbidimetry method [LASAY Lp(a) auto; SHIMA Laboratories Co., Ltd] with a normal value of <30 mg/dL. An Lp(a) protein validated standard was used to calibrate the examination, and the coefficient of variation value of repetitive measurements was <10%. The concentrations of glucose were measured by enzymatic hexokinase method. HbA_1c was measured using a Tosoh Automated Glycohemoglobin Analyzer HLC-723G8. Concentrations of fibrinogen were measured using a Stago auto analyzer by the Clauss method with an STA Fibrinogen kit (Diagnostica Stago, Taverny, France). Big endothelin-1 (ET-1) was measured using a highly sensitive and specific commercial sandwich enzyme immunoassay (BI-20082H; Biomedica, Wien, Austria). Plasma N-terminal pro–B-type natriuretic peptide (NT-proBNP) was determined using an electrochemiluminescence immunoassay (ECLIA) method (Roche Diagnostics, Mannheim, Germany) with a Roche modular analytics E170 immunoassay analyzer.

Angiographic data were evaluated from catheter laboratory records by three experienced interventional cardiologists in accord with our previous studies (18). The Gensini score (GS) was calculated as the stenosis score multiplied by the location score for all diseased segments. The severity score of each coronary lesion was defined according to the narrowing degree of the coronary artery and its importance (19).

The values were expressed as the mean ± SD or median (25th–75th percentile) for the continuous variables and the number (percentage) for the categorical variables. The Kolmogorov-Smirnov test was used to test the distribution pattern. The differences in clinical characteristics between groups were analyzed using Student t test, Mann-Whitney U test, χ² tests, or Fisher exact test where appropriate. The event-free survival rates among groups were estimated by the Kaplan-Meier method and compared by the log-rank test. Univariate and multivariate Cox regression analyses were performed to calculate the hazard ratios (HRs). A P value <0.05 was considered statistically significant. The statistical analyses were performed with SPSS, version 21.0, software (SPSS, Chicago, IL) and R language, version 3.5.2 (Feather Spray).

As presented in Fig. 1, among 5,143 subjects, 18.8%, 43.5%, and 37.6% were diagnosed as having NGR, pre-DM, and DM, respectively, according to ADA criteria. The baseline characteristics of the study participants were shown in Table 1. The age, BMI, glucose, HbA_1c, TG, and fibrinogen were positively associated with the status of glucose metabolism from NGR to DM (all P < 0.001). The percentage of male patients decreased, while the proportion of patients with hypertension was elevated from NGR to DM (P < 0.001). Patients with pre-DM and DM had higher levels of TC and NT-proBNP than the NGR group. Meanwhile, DM but not pre-DM patients had significantly lower levels of HDL-C and left ventricle ejection fraction (LVEF) but higher levels of big ET-1 than the NGR population. There was no significant difference regarding smoking, family history of CAD, creatinine, and medication prescriptions among the three groups (P > 0.05).

Over a median follow-up time of 6.1 years (25th–75th percentile 5.6–6.8), 435 CVEs occurred (154 died, 76 suffered nonfatal MI, and 205 had nonfatal strokes). The prevalence of CVEs in the NGR, pre-DM, and DM groups was 5.8%, 7.7%, and 10.7%, respectively. Kaplan-Meier analysis (Fig. 1A) showed that DM subjects had the lowest event-free survival rate among the three groups (P < 0.05), while there was no significant difference between that of pre-DM and NGR groups (P > 0.05). In the current study, Lp(a) levels of the population had skewed distribution (Supplementary Fig. 2), which was consistent with previous studies. By Lp(a) status (Lp(a) <10 mg/dL, 10 ≤ Lp(a) < 30 mg/dL, 30 ≤ Lp(a) < 50 mg/dL, and Lp(a) ≥50 mg/dL), patients with Lp(a) ≥50 mg/dL were least likely to be free of events (Fig. 1B). However, when the patients were evaluated according to both glucose metabolism and Lp(a) status, those with Lp(a) ≥50 mg/dL had significantly higher risk of CVEs than the reference group (NGR plus Lp(a) <10 mg/dL) in all glucose metabolism status. Meanwhile, DM plus 10 ≤ Lp(a) < 30 mg/dL, pre-DM plus 30 ≤ Lp(a) < 50 mg/dL, and DM plus 30 ≤ Lp(a) < 50 mg/dL groups had significantly lower cumulative event-free survival rates compared with the reference group [NGR plus Lp(a) <10 mg/dL group] (Fig. 1C) (all P < 0.05, respectively).

As presented in Supplementary Table 1, univariate Cox regression models showed that patients with DM had 1.860-fold higher risk of CVEs than NGR subjects (HR 1.860 [95% CI 1.385–2.499], P < 0.05). Additional adjustment for other variables did not change the significance of association. The pre-DM group did not have an increase in CVE risk compared with the NGR group (P > 0.05). Moreover, per 1-SD change in Lp(a) (26.5 mg/dL in pre-DM and 26.0 mg/dL in DM) was associated with 32.7% and 38.6% increased risk of CVEs in pre-DM and DM, respectively (Table 2). Multivariate Cox regression analyses according to both glucose metabolism and Lp(a) status indicated that patients in DM plus 10 ≤ Lp(a) < 30 mg/dL, pre-DM plus 30 ≤ Lp(a) < 50 mg/dL, and DM plus 30 ≤ Lp(a) < 50 mg/dL groups had 2.594-fold (95% CI 1.396–4.820), 2.181-fold (95% CI 1.099–4.327), and 3.088-fold (95% CI 1.535–5.895) higher risk of CVEs (Table 2) (all P < 0.05, respectively). Patients with Lp(a) >50 mg/dL was associated with 2.369-, 2.668-, and 3.470-fold risk of CVEs in NGR, pre-DM, and DM groups, respectively.

In Cox prediction models, with traditional risk factors, C-statistic values were 0.770 (95% CI 0.768–0.854), 0.700 (0.651–0.750), and 0.675 (0.630–0.719) for NGR, pre-DM, and DM, respectively (Table 3). Addition of Lp(a) to the original model resulted in a significant improvement in C-statistic for pre-DM and DM patients (∆C-statistic 0.022 [0.003–0.046], P = 0.043, and ∆C-statistic 0.029 [0.013–0.048]), P = 0.001) but not for those with NGR (∆C-statistic −0.001 [−0.022–0.009], P = 0.876).

Plasma Lp(a) was reported to be associated with cardiovascular risk in many prospective cohort studies (7,8). It was reported that individuals with diabetes showed the highest Lp(a)-associated risk, while no significant difference of the Lp(a)-associated risk was identified when patients were stratified by other high risk states (13). In this study, we investigated the impact of high Lp(a) on CVEs in stable, angiography-proven CAD patients with different glucose metabolism status. Cox regression analysis indicated that patients with DM but not those with pre-DM had higher risk of CVEs when the patients were simply divided into the three groups: DM, pre-DM, and NGR. Interestingly, when patients were categorized according to both status of glucose metabolism and Lp(a) levels, patient with pre-DM plus 30 ≤ Lp(a) < 50 mg/dL and patients with pre-DM plus Lp(a) ≥50 mg/dL had 2.181-fold and 2.668-fold increased risk of CVEs, respectively, compared with that in subjects with NGR and Lp(a) <10 mg/dL. Moreover, adding Lp(a) to the model improved the risk prediction for CVEs in pre-DM and DM groups but not in NGR. Thus, our study, for the first time, suggested that patients with high Lp(a) and pre-DM had high risk for CVEs.

High Lp(a) is a lipid abnormality mainly (90%) caused by variation in the LPA gene. Circulating Lp(a) levels are rarely influenced by dietary and environmental factors; thus, high Lp(a) levels might have lifelong effects on the health of individuals (20,21). In cross-sectional studies, plasma Lp(a) was proved to be associated with coronary calcification and extent of coronary stenosis (15,22). Interestingly, Lp(a)-hyperlipoproteinemia also increased the risk of early-onset CAD for patients with clinical familial hypocholesteremia (16). Additionally, most studies about Lp(a) in CVD risk prediction have had positive results (5,13,14). For example, the Copenhagen City Heart Study showed that individuals with Lp(a) concentrations in the 67th–90th percentile of the population had a 1.6-fold increased risk for incident MI compared with those with Lp(a) <5 mg/dL (5). Results from the BiomarCaRE consortium also indicated that individuals with Lp(a) levels ≥90th percentile had 1.44-fold increased risk for CVD compared with risk in those with Lp(a) levels in the lowest third (13). However, these studies mostly involved patients who were free from CAD at baseline. Of note, inconsistent results were seen for patients with established CAD. In a study of patients with recent acute coronary syndrome and receiving statin therapy, Lp(a) was not associated with CVEs (23). Therefore, more studies concerning the prognosis of Lp(a) in patients with CAD at baseline might be needed.

Currently, according to the European Society of Cardiology/European Atherosclerosis Society (ESC/EAS), measurement of Lp(a) is recommended in patients with intermediate or high risk of CVD, particularly in those with premature CVD, familial hypercholesterolemia, a family history of premature CVD, and/or elevated Lp(a) (24). Canadian Cardiovascular Society (CCS) Guidelines for the Management of Dyslipidemia recommended adding Lp(a) in the further risk assessment when subjects were with medium CVD risk according to Framingham Risk Score or when they were with a family history of premature CAD (25). American College of Cardiology (ACC)/American Heart Association (AHA) guidelines did not give recommendations about Lp(a) measurement (11). Besides, a novel therapeutic agent, IONIS-APO(a)Rx, an antisense oligonucleotide targeting hepatic apo(a) mRNA, had been evaluated and proved to be tolerable and effective (26). In this situation, more specific information on the predictive value of Lp(a) in different subgroups and populations is valuable. Previous meta-analysis reported that Lp(a) was linearly related with CVD risk in statin-treated patients (27). Results from the BiomarCaRE consortium indicated that HRs for Lp(a) were comparable across subgroups defined by age, sex, smoking, hypertension, BMI, regions, and LDL-C level but individuals with diabetes were with significantly higher risk (13). However, no current guidelines recommended measurements or treatments of high Lp(a) levels in patients with impaired glucose metabolism.

Approximately 415 million individuals have DM worldwide and almost three times as many patients have pre-DM (28). Strong evidence has indicated that CAD is a common comorbidity in patients with pre-DM and DM (29). Our study shows that DM but not pre-DM is positively associated with CVEs after adjustment for confounding factors. According to studies conducted by Liu et al. (18) and Qiu et al. (30), patients with pre-DM are more likely to have higher CVE risk when combined with other disorders such as hypertension. It has also been demonstrated that Lp(a) is independently associated with the severity of CAD in patients with T2DM but not in those with NGR (31). Indeed, Saeed et al. reported that elevated Lp(a) levels had a predictive value in CVD risk in apparently healthy Caucasian individuals with DM or pre-DM (14). In the current study, we not only analyzed the prognosis of Lp(a) separately within NGR, pre-DM, and DM but also gave special attention to the impacts of high Lp(a) plus different glucose metabolic status. As the main novel findings of our study, patients with both Lp(a) >50 mg/dL and DM had 3.470-fold higher risk of CVEs. Meanwhile, pre-DM plus Lp(a) >30 mg/dL or 30 ≤ Lp(a) < 50 mg/dL but not pre-DM alone had worse cardiovascular outcomes, indicating clinical importance of Lp(a) measurement and intervention in patients with impaired glucose metabolism.

Nevertheless, the current study had several limitations. Firstly, this is a study among Chinese patients with stable CAD. However, considering the high incidence and mortality of CAD in patients with impaired glucose metabolism worldwide, our results may also be applicable to other populations. Secondly, we measured Lp(a) only at baseline, and the follow-up levels of Lp(a) may also be clinically significant. Thirdly, we did not assess all of the metabolic factors and parameters of insulin resistance due to the features of patients in our study. More study is needed to confirm our findings due to the moderate sample size and limited follow-up time of the current study.

In conclusion, in our multicenter study, data, for the first time, indicated that the pre-DM patients who also had high Lp(a) had worse prognosis. Moreover, patients with DM and high Lp(a) were with highest risk of CVEs, suggesting that measurement of Lp(a) and treating strategies toward Lp(a)-hyperlipoproteinemia might be beneficial in pre-DM and DM patients.

Acknowledgments. The authors thank all the staff and participants of this study for their important contributions.

Funding. This work was partially supported by the Capital Health Development Fund (201614035) and CAMS Major Collaborative Innovation Project (2016-I2M-1-011) awarded to J.-J.L.

The study sponsors did not participate in the study design; the collection, analysis, or interpretation of data; the writing of the manuscript; or the decision to submit the manuscript for publication.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. J.-L.J. completed the project, analyzed data, and wrote the manuscript. Y.X.-C., H.-W.Z., and D.S. contributed to data collection. Q.H. and Y.-F.L. contributed to the collection of data. Y.-L.G., N.-Q.W., and C.-G.Z. contributed to recruitment of patients, clinical diagnosis of disease, and data collection. Y.G., Q.-T.D., H.-H.L., and Q.D. contributed to the collection of clinical data and procedure of laboratory examination. J.-J.L. designed the study, interpreted data, and contributed to critically revising the manuscript. All authors approved the final article. J.-J.L. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.